Skip to contents

The manifest for a series is a plain text file that lists the files contained within the series.


manifest_read(series, origin = bucket_remote_url())

manifest_write(series, date = Sys.Date(), origin = bucket_local_path())

manifest_build(series, date, origin = bucket_local_path())



Name of the series (e.g., "series-rosemary")


Location in which to find the series


Publication date for the series (defaults to current date)


Tibble containing the manifest data, returned visibly to the user by

manifest_read() and manifest_build(), and invisibly by

manifest_write(). The tibble contains one row per image file in the series, and the following columns:

  • series_name is a string with the series name (e.g., "series-rosemary")

  • series_date contains the publication date in YYYY-MM-DD format

  • path specifies the path to the image (e.g., "800/rosemary_001_1000_short-title.png")

  • folder specifies the directory part of the path (e.g., "800")

  • file_name specifies the file name part of the path (e.g., "rosemary_001_1000_short-title.png")

  • file_format specifies the file format for the image (e.g., "png")

  • system_name specifies the generative art system name (e.g., "rosemary")

  • system_version specifies the generative art system version identifier (e.g., "001")

  • image_id specifies the identifier string for the image (e.g., "1000")

  • image_short_title specifies the short title for the image, or NA (e.g., "short-title")

  • image_long_title specifies the long title for the image, or NA (e.g., "The Long Title")

  • manifest_version specifies the version of the manifest format used (currently always 1)

For manifest_read() and manifest_build() returning this tibble is the only thing it does. For manifest_write() the tibble is written to a csv file whose location is specified using the

destination argument.


The manifest file is used to document the content of a published art series, and manifest_*() functions provide tools to work with manifests with a minimum of pain. The typical workflow is expected to be as follows. Once an art series has been finalised and all images have been written into the "local bucket folder", use manifest_write() to create a manifest file within the local series folder. Once that is done, you can upload the completed series to the "remote bucket folder".

Whenever you need to inspect the contents of the now-online series, read the manifest file from the remote bucket using manifest_read(). As an example, you can use this to programmatically construct an HTML document that displays all images in a series, by calling manifest_read() within a code chunk in a quarto or R markdown document.

This workflow is one in which manifests are constructed locally, published to a remote, and then read from the remote location. For that reason, the default behaviour is that manifest_read() sets the origin to bucket_remote_url(), whereas manifest_write() sets the origin and destination to bucket_local_path().

In some cases it may be convenient to construct the manifest tibble from the series image files without writing it to a csv file. To that end there is also a manifest_build() function that does this.